Diffusion360_ComfyUI

★ 45

全景生成360度全景文本到全景图像到全景

通过文本或透视图生成无缝360°全景图，内置 text-to-pano 与 image-to-pano 流程，便于在 ComfyUI 中快速制作环景。

💡 将文本描述或照片一键转为无缝360°全景图用于虚拟漫游。

🍴 9 Forks💻 Python🔄 2025-03-17

🔗 GitHub 原文

📦

网盘下载

复制链接后前往夸克网盘下载

https://pan.quark.cn/s/79aaff81621b

📦 requirements.txt

torch
torchvision
torchaudio
diffusers==0.26.0
accelerate
xformers
triton
transformers
realesrgan==0.3.0
py360convert
numpy==1.23.5

📄 README

Diffusion360_ComfyUI

ComfyUI plugin of https://github.com/ArcherFMY/SD-T2I-360PanoImage

features

[x] text-to-panoimage pipeline

[x] image-to-panoimage pipeline

Installation

Install ComfyUI following https://github.com/comfyanonymous/ComfyUI

Install this plugin by the following commands.

“`

cd ComfyUI/custom_nodes

git clone https://github.com/ArcherFMY/Diffusion360_ComfyUI

cd Diffusion360_ComfyUI

pip install -r requirements.txt

“`

__Note that I recommend to install 0.20.0<= diffusers <= 0.26.0. The higher diffusers version will get an over-saturated SR result.__

Usage

prepare the models

Download models from HF. Rename the folder sd-t2i-360panoimage to diffusion360 and put the folder into models/diffusers.

${ComfyUI_ROOT} 
|-- models
    |-- diffusers
        |-- diffusion360
            |-- sd-base
            |-- sr-base
            |-- sr-control
            |-- sd-i2p
            |-- RealESRGAN_x2plus.pth
            |...

load workflows

Load workflows from the folder workflows/

${This_Repo_ROOT} 
|-- workflows
    |-- Text2Pano.json ## text-to-pano pipeline
    |-- Image2Pano_load_image.json ## image-to-pano pipeline, load a square image from local.
    |-- Image2Pano_gen_image.json ## text-to-pano pipeline, generate a square image from a certain SD model. (only support batchsize=1)